In a recent Croakey post, public health researcher Professor Chris Rissel reported on new research which found that one in five adults say they would cycle more if they didnâ€™t have to wear a bicycle helmet.
Tim Churches, a Sydney-based epidemiologist with a personal interest in active transport and urban re-design who has previously critiqued work by Rissel, has had a close look at the study, and disagrees with some of its methodology and conclusions.
Rather than arguing for the end of mandated cycle helmet wearing, Churches says it would be better for Australian governments and to focus on strategies which have been proven to work in Europe, such as providing separated cycle lanes on busy roads, reducingÂ speed limits on residential and minor roads, and reforming road rules and accident liability laws to make motor vehicle drivers more responsible for cyclist and pedestrian safety.
Letâ€™s focus on proven interventions to promote cycling
Tim Churches writes:
In a recent Croakey article titled, More people would cycle if helmets were not compulsory: new study, Professor Chris Rissel reports on the results of a population-based survey which he and Dr Li Ming Wen commissioned into attitudes to compulsory cycling helmet laws amongst Sydney residents.
In the Croakey article, Professor Rissel makes the following claim, based on the study survey results:
“While a hypothetical situation, if only half or a quarter of the one in five respondents who said they would cycle more if they did not have to wear a helmet did ride more, Sydney targets for increasing cycling would be achieved by repealing mandatory bicycle helmet legislation. This increase in cycling would result without having to spend millions of dollars on new cycling infrastructure.”
He makes similar claims in an article in “The Conversation” on the same study, titled “Make helmets optional to double the number of cyclists in Australia”.
These are fairly startling claims which warrant careful examination, particularly when they are presented as offering government authorities the opportunity to meet cycling participation targets without, as Professor Rissel puts it, “having to spend millions of dollars on new cycling infrastructure”.
The cycling participation targets referred to are discussed in the full paper on the study, published in the December 2011 issue of the Health Promotion Journal of Australia. The primary target is contained in the Australian National Cycling Strategy 2011-2016, which calls for a doubling of the number of people cycling by 2016. It proposes three measures of cycling participation: the proportion of the population who have cycled at some time in the last week, in the last month, and in the last year.
The other target discussed is given in the 2010 NSW State Plan, which called for an increase to 5% in the cycling mode-share for all trips made at the local and district level in the Greater Sydney region (from the current Sydney cycling mode-share of about 1.2%). This target has since been superseded in the NSW 2021 plan by a more modest target of a doubling of local trip cycling mode-share in Sydney by 2021.
Professor Rissel provides the following justification for his claims (he gives a similar justification in his “The Conversation” article; text in square brackets has been added to provide context):
“There are about 3.5 million people in Sydney aged 16 years and older. Conservatively 60% havenâ€™t ridden a bike in the past year â€“ leaving 2,100,000. With 19% of people not having ridden [in] the past year saying they would ride more [if they were not compelled to wear a helmet], this represents 399,000 potential riders [if the mandatory helmet requirement were removed].”
So far, so good: the 60% not having ridden in the last year is derived from the study survey results, and this figure is comparable with that found in the much larger 2011 Australian Cycling Participation survey, and the 19% figure is similarly based on the Wen and Rissel survey.
Rissel advises in “The Conversation” article that it is reasonable to “…halve the number of people saying they’d ride but don’t (even the best intentions aren’t always followed through)…”.
In the full paper, as in his Croakey article, he suggests a good-intentions to actual-behavioural-change correction factor of between one-quarter and one-half. Therefore, based on their survey results and Rissel’s own advice, a reasonable estimate of the hypothetical increase in the at-least-once-in-the-last-year adult cyclist numbers in Sydney, if the mandatory helmet road rule were to be removed, would be between 100,000 and 200,000, with an upper bound of about 400,000. Thus the estimate of the total number of at-least-once-in-the-last-year adult cyclists in Sydney would be about 200,000 new (presumably unhelmeted) cyclists plus about 1.2 million existing at-least-once-a-year adult cyclists (based on the Wen and Rissel survey estimate that 34% of the adult Sydney population of about 3.5 million reporting had cycled in the previous year), giving a hypothetical total of 1.4 million cyclists.
The correct baseline with which to compare this hypothetical no-mandatory-helmets estimate is, of course, the estimate of current at-least-once-in-the-last-year adult cyclists from the same survey: about 1.2 million.
Thus, based on their survey results, the hypothetical increase in the at-least-once-in-the-last-year cycling participation in Sydney adults, if the compulsory helmet requirement were to be removed, would be an increase from 1.2 million to 1.4 million people, which is a 17% change.
This is a great deal less than the 100% increase required by the 2016 cycling participation target.
However, in justifying his claims for meeting cycling targets, Professor Rissel inexplicably does not use the estimate of current adult cycling participation in Sydney from his own survey. Instead, he derives a baseline cycling participation number for use in his calculations as follows:
“In Sydney, the  Census tells us that about 10,000 people rode to work on Census day.Â We know that this is an under-estimate of cycling levels, but even if we multiply this by a factor of 10 this 100,000 Sydney cyclists are still a quarter of the potential new cyclists.”
In his “The Conversation” article, he justifies his claims thus:
“Compare this figure [the hypothetical 200,000 additional cyclists if helmets were optional] to the 10,000 people who ride to work on Ride to Work Day. Even if you multiply this group by ten to include the 10% of the population who occasionally ride, and then halve the number of people saying theyâ€™d ride but donâ€™t (even best intentions arenâ€™t always followed through), this would still double the number of people currently cycling in Sydney.”
To use these figures, based on the 10,000 people who rode to work on Census day in August 2006 (or the apparently identical number of people who commuted by bicycle on Ride to Work Day), multiplied by an arbitrary and unreferenced factor of 10, as an estimate of current at-least-once-in-the-last-year cycling participation is just not valid, and cannot possibly be justified when a far more direct and recent estimate of current at-least-once-in-the-last-year cycling participation is available from the authors’ very own study survey.
This objection to Professor Rissel’s justification for his claims is not just an arcane quibble: his claims are based on a hypothetical before-and-after intervention study, where the intervention is the removal of the mandatory bike helmet requirement and the outcome is the predicted change in cycling participation by Sydney adults, informed by the findings of the authors’ research survey.
A basic epidemiological principle of any before-and-after intervention study, even one conducted as a thought experiment such as this, is that the method used to measure or estimate the outcome (cycling participation in this case) must be the same both before and after the intervention – otherwise it would be unclear whether any apparent change were due to the intervention or to differences in the measurement or estimation method.
It is not scientifically acceptable for Professor Rissel to substitute a “before” estimate derived in a completely different (and questionable) manner in his calculations, particularly when a perfectly valid “before” estimate is available in his own survey results.
Other criticisms of the Rissel and Wen study by Dr Alan Davies can be found on his Melbourne Urbanist blog.
Thus, in short, the claim made by Professor Rissel that the removal of the mandatory helmet road rule would result in the 2016 cycling participation targets being met is not supported by the results of his own study.
Now, it could be argued that, of the people who have not cycled in the last year and who indicated that they would cycle more often if the mandatory helmet rule were removed, a substantial proportion might become regular cyclists, perhaps riding even once a month or more.
However, even if we assume that all of the estimated 200,000 people who might hypothetically increase their cycling if they didn’t have to wear helmets took up cycling at least once a month, then the increase in the at-least-in-the-last-month cycling participation rate would still be only 32%.
Of course, such an assumption is unrealistic, and the increase in at-least-once-a-month cyclists would most likely to be substantially less. In other words, no matter how the hypothetical increase in adult cycling participation is estimated, the claim by Professor Rissel that removing the mandatory helmet requirement would result in cycling participation targets being met is just not supported by his own data.
In order to substantially increase cycling participation and mode-share, and to meet agreed targets, Australian governments and authorities need to focus on strategies which have been proven to work in Europe.
These include provision of separated cycle lanes on busy roads, the reduction of speed limits on residential and minor roads, and reform of road rules and accident liability laws to make motor vehicle drivers more responsible for the safety of vulnerable road users such as cyclists and pedestrians.
Some of these measures require capital investment, although the level of funding required is tiny when compared to the cost of major road infrastructure projects for motor vehicles.
Others, such as lowering residential street speed limits, may save money due to lower accident and injury costs, and may help reclaim our often deserted neighbourhood streets for all forms of active transport, not just cycling.
Rescinding the two-decade-old regulations that require helmets be worn when riding on the road is not a viable alternative to such real action, if even our very modest cycling targets are to be met.